hashicorp/packer · error

can't compute sum of opposing infinities

Error message

can't compute sum of opposing infinities

What it means

The `sum()` HCL2 function recovers a big.Float ErrNaN panic caused by adding +Inf and -Inf (or Inf - Inf), which is undefined math, and converts it into this regular error to stay inside the cty Function abstraction.

Source

Thrown at hcl2template/function/sum.go:52

		arg := args[0].AsValueSlice()
		ty := args[0].Type()

		if !ty.IsListType() && !ty.IsSetType() && !ty.IsTupleType() {
			return cty.NilVal, function.NewArgErrorf(0, "argument must be list, set, or tuple. Received %s", ty.FriendlyName())
		}

		if !args[0].IsWhollyKnown() {
			return cty.UnknownVal(cty.Number), nil
		}

		// big.Float.Add can panic if the input values are opposing infinities,
		// so we must catch that here in order to remain within
		// the cty Function abstraction.
		defer func() {
			if r := recover(); r != nil {
				if _, ok := r.(big.ErrNaN); ok {
					ret = cty.NilVal
					err = fmt.Errorf("can't compute sum of opposing infinities")
				} else {
					// not a panic we recognize
					panic(r)
				}
			}
		}()

		s := arg[0]
		if s.IsNull() {
			return cty.NilVal, function.NewArgErrorf(0, "argument must be list, set, or tuple of number values")
		}
		s, err = convert.Convert(s, cty.Number)
		if err != nil {
			return cty.NilVal, function.NewArgErrorf(0, "argument must be list, set, or tuple of number values")
		}
		for _, v := range arg[1:] {
			if v.IsNull() {
				return cty.NilVal, function.NewArgErrorf(0, "argument must be list, set, or tuple of number values")

View on GitHub (pinned to eb36e3c3e4)

Solutions

1. Remove or fix the offending values so +Inf and -Inf are not both present
2. Sanitize the list before summing: filter or replace infinities with finite bounds
3. Check upstream math (divisions by zero) that produce infinities
4. Example: use `can()`/`try()` to drop non-finite elements

Example fix

// before
total = sum([1, pow(0, -1), -pow(0, -1)])  // +Inf and -Inf -> error
// after
nums = [for n in raw_nums : n if !isnan(n) && !isinfinite(n)]
total = sum(nums)

// filter non-finite values before summing
locals {
  finite = [for n in var.nums : n if n != pow(0, -1) && n != -pow(0, -1)]
  total  = sum(local.finite)
}

// tolerate and fall back
locals {
  total = try(sum(var.nums), 0)
}

When it happens

Trigger: `sum([...])` over a list of numbers containing both +Inf and -Inf (e.g. numbers constructed via `pow(0, -1)`-style infinities or from computed data).

Common situations: Building number lists from division expressions that can yield infinities; passing computed floats from data sources whose math produced Inf on both signs.

Related errors

• too many values, 1 needed: %v
• failed to write gzip raw data: %w
• failed to flush gzip writer: %w
• failed to close gzip writer: %w
• failed to read file %q: %s

AI-assisted analysis of hashicorp/packer@eb36e3c3e4 (2026-09-05). Data as JSON: /api/errors/2b308c659c08e8cf. Report an issue: GitHub.